[gecode-users] Copying spaces

[Christian Schulte]

That is absolutely bizarre! unforce de-registers a propagator that requires
deletion (many propagators do not have to be deleted if their space are
deleted. some need explicit deletion as they keep external memory for
example).

So, I have no idea, I thought for a while. Do you use anything self-built
(propagator, branching)? Or maybe is your copy function and/or the
constructor called by the copy function wrong (forgot something to update
and just assigned it)? That is most likely.

Christian

--
Christian Schulte, www.ict.kth.se/~cschulte/


-----Original Message-----
From: users-bounces at gecode.org [mailto:users-bounces at gecode.org] On Behalf
Of Malcolm Ryan
Sent: Wednesday, February 04, 2009 10:03 AM
To: gecode list
Subject: Re: [gecode-users] Copying spaces

On 04/02/2009, at 7:14 PM, Christian Schulte wrote:
> Copy() is an internal function (maybe we should make it protected)  
> use by
> the clone() operation. So you have to use clone() to create a clone  
> of a
> space!

Thanks, that sorted out that problem... but left me with another. I  
haven't quite isolated it down to a simple example yet, but can you  
offer any idea what might be causing this:

Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x0033d000
Gecode::Actor::unforce () at core.icc:1591
1591          while (this != *f)

The GDB backtrace shows:

#0  Gecode::Actor::unforce () at core.icc:1591
#1  Gecode::Actor::unforce () at core.icc:1591
#2  Gecode::Int::Extensional::Base<Gecode::Int::IntView,  
true>::dispose (this=0x1031038, home=0xf23d20) at gecode/int/ 
extensional.cc:145
#3  0x0058ae7d in Gecode::Actor::unforce () at core.icc:1646
#4  0x0058ae7d in  
Gecode::Int::Extensional::Basic<Gecode::Int::IntView>::propagate  
(this=0x1031038, home=0xf23d20) at gecode/int/extensional.cc:145
#5  0x0026a043 in Gecode::Space::status (this=0xf23d20, pn=@0x27c1b8)  
at gecode/int/extensional.cc:145

It occurs at a point in the code where I have just cloned a space and  
committed a Branching. Calling status() before the commit is fine.  
Calling it again after the commit causes this crash. If I run the same  
code on the original space (not the clone) it works fine.

I know that this is not really enough detail, but if you could give me  
any idea of what kind of problem I might be facing, it could help me  
track down the bug. What is the unforce() function doing at this point?

Malcolm 
_______________________________________________
Gecode users mailing list
users at gecode.org
https://www.gecode.org/mailman/listinfo/gecode-users